Abrssive wesr resistqnce of qustempered ductile ironql room lemperclureJ'M. SCHISSLER, P. BRENOT and J.P. CHOBAUT, Unité de Science et Génie des Matériaux Métalliques, U.S.G.2.M., UA 159, parc deSaurupt, 54042 Nancy Cedex, France.

AbstractThe use of austempered ductile iron obtained by heat treating spheroidal-graphite cast iron is very important in the gear industry. However thenature of the bainitic structure, which is used to improve weàr resistancel has never been f ully identif ied.The aims of thts study were:- to optimize the upper batnitic structures in S.G. cast irons (usually called austempered ductile iron or A.D.l.) to have the best abrasive wear

resistance at room temperature.to study the influence of the solidification cell characteristics upon these optimized structures.

- to determine the evolution of the microstructures during the abrasion test.Results clearly indicate the primordial influence of the retained austenitic phase. A high percentage of retained austenite promotes a highabrasive wear resistance. The analysis of the results shows that this austenitic phasé is heterogóneous. The best abrasive wear resistince isassociated with the lowest hardness value. During abrasion, austenite at or near the surface is gradually and partly transformed to martensite.

RiassuntoResistenza all'usura abrasiva della ghisa sferoidale, sottoposta ad austempering, a temperatura ambienteL'uso di ghisa con grafite sferoidale austemprata sta diventando molio lmportante nell'industriàdegli ingranaggi. Ciò nonostante iltrattamentotermico bainitico, che è necessario per migliorare la resistenza all'usura, non è mai stato ben identificato.ll presente studio si proponeva di:- ottimizzarelestrutturebainitichesuperiori nelleghisesferoidali (spessochiamate"austemperedductileiron"o"A.D.l.")misurandonela

resistenza all'usura abrasiva, a temperatura ambiente, in una macchina che simula l'abrasione.- studiare l'influenza delle caratteristiche delle celle dì solidificazione su tali strutture ottimizzate.- determinare l'evoluzione delle microstrutture durante la prova di abrasione.I risultati indicano chiaramente l'influenza fondamentale della fase austenite residua la quale, se in alta percentuale, favorisce una più altaresistenza all'usura abrasiva. L'analisi dei risultati mostra che tale fase è eterogenea. La resistenza più elevata all'usura abrasiva e associata aivalori di durezza più bassi. Durante l'abrasione, l'austenite in superficie o presòo la superficìe si trasiorma gradualmente e parzialmente inmartensite.

lntroduction

Austempered ductile irons (A. D. l.) obtained byisothermal heat treatment of spheroidal graphite castirons (S.G. cast irons) are known to produce beneficialmechanical properties. However, these properties

Fig. 1 - Diagram of the abrasive wear test machine

1 : electric motor2 : cylindrical test specimens3 : silicon carbide4 : heat element

depend strongly on the composition, and heattreatment conditions (austenitisation and isothermalreactions) ('l-7).Some authors have shown that A.D.l. structures have a

lower wear rate than does a quenched and temperedsteel of similar hardness (8-9). To our knowledge, theinfluence of different A.D.l. microstructures on wearresistance has never been studied before, and the aimof this work is to determine this correlation.

Experimental procedures

ln this study an unalloyed ducttle iron was used, thechemical composrtton of which was 3.40% C; 2.46%Si; 0.08% Mn; 0.008% S;^0.05% Mg. This alloy had acount of 200 nodules lmm'. Cylindrical test specimens(L : B0 mm, R : 6 mm) were austempered for varioustimes to produce a variety of microstructures. Thefollowing heat treatment cycles were used:- austenitise at 930'C for t h.

- austemper in a wood's alloy bath at 380'C for varioustimes ranging from 5 min to 1200 min.

- water quench.A.D.l. specimens were tested in an abrasion testdevice described in previous publications (10). Themachine, sketched in Fig. 1, stirs the specimensthrough abrasive material (silicon carbide - 2 mm" granulometry "). The characteristics of this machineare:

- rotation speed 200 rpm (linear speed: 2.5 m/sec-r atthe end of the specimen).

- test temperature: room temperature.Hardness measurements, made at room temperature,were HV 30 type; magnetic measurements were madeusing a Magne-Gage apparatus.

Vol 5 t3l (1987)

Microstructure observations

Various microstructures were observed byconventional optical microscopy after nital etching andthe abraded surfaces were examined by scanningelectron microscopy (SEM). Phase percentages weredetermined using a Texture Analyser System (T.A.S.)marketed by E. Leitz A.G. and the apparent degree ofadvancement of the reaction was analysed bydilatometry.The loss of matter, AP, was measured every 0.5 millionrevolutions and compared with the initialweight of thespecimen. Wear rate was defined as the ratio AP/P.

Fig. 2 - Variation of the first stage bainitic transformation as a function o1

austempered time at 380'C

5 I 10 12 1s 20 3omin Holding at 380'C

Fig. 3 - Variations of wear rates as a function of austempering time at 380"C

^P/P.10{20

18

16

14

12

10

I

6

4

2

Holding time at 380'C

o -5 min* -20 hours

. -30 min

a -10 min

' -15 min

t (Hours)0

Results and discussion

The first tests were made to determine how to obtaindifferent microstructures. Specimens were cooledfrom 930'C to 380"C and held for various times. Thepercentage austenite transformed to bainite is given inFig. 2. lt can be seen that the transformation is 90%complete in 20 min. ln order to obtain a well finishedbainitic product, one specimen was heat treated for1200 min (20 h) at 380'C. Wear tests were run at roomtemperature on these heat-treated specimens.Fig. 3 illustrates the variation in wear rate as a functionof austempering time. A comparison between Fig. 2and Fig. 3 shows that the best resistance to wear doesnot correspond to the end of the bainitic reaction. lnfact, the highest wear resistance is associated with amicrostructure obtained after 10 or 15 min bainiticholding time at 380"C (about B0% bainite).Fig. 4 is most instructive, because it shows that thesetimes correspond to the maximum of retainedaustenite and to the lower values of both the HV 30hardness and the magnetic characteristic o. Other veryimportant results, such as the following, can be noticedfrom Fig.4:o Even after 20 h at 380'C, the microstructure

contains austenite, which means that the bainiticreaction is not finished, and this in turn explains theapparent end of the reaction noticed in Fig. 2.

o Again, the best abrasion resistance is noùssociatedwith the highest hardness value.

Fig. 4 - Evolutions of characteristics such as:* hardness -a* magnetic property - b.* retained austenite percentage - c -

Hardness HV30

380370

360

350340

10-6Tesla .m3.kg-1

125

115

105

o9585

75

65

Retained austenite Vo

40

30

20

10

0r2oo min Holding time at 380 'C5 1015 30

Metallurgical Science and Technology

These important remarks illustrate the importance ofthe upper bainitic reaction. The fact that ductile ironsare cast alloys increases the complexity of the problem.It must be remembered that the upper'bainitic reaction,occurring in Fe-C-Si alloys, such as homogeneoussteels, can be divided into two stages (j1-,12)y--+ (c) + (y): lststage(c) + 1y1-> u, * carbides: 2nd stagewhere y : initialaustenite

(a) : carbon supersaturated ferrite(y): carbon enriched austenite,,, post-bainitic

austenite "a, : ferrite

Carbides produced during the 2nd stage are induced bythe desaturation of (a) and initiated bylhedecomposition of (y). The enriched austenite (y) has anew Ms temperature and therefore is generally stablef rom room temperature to - 130.C evòn though the Mstemperature of the initial austenite is above 2Ò.C.These two stages are well separated.The evolution of the retained austenite percentagewitnesses to the existence of several kinds ofaustenite. This evolution is the result of both reactionsy + (a) + (y) and y --+ o, * y, the second reactionggggrr1rrg during the quenching operation from 390.C ro20'C. The retained austenitic phase contains both y and(y)which are the initialand the enriched phases. Underthese conditions it is clear that the amount of austenitereaches a maximum value but this peak is notassociated with the maximum enriched austenitepercentage.ln ductile irons, which are Fe-C-Si cast alloys, theproblem is more complex because the initial structureconsists of solidification cells, and each of themcontains a graphite nodule (Fig. S)

Fig. 5 - Microstructure ofthe S.G. cast iron in its as-cast state: the solidification cellboundary (Zone lll) contains pearlite

Vot. 5 [s](1987)

The chemical composition of this cell is mostheterogeneous. Usually the matrix inside a single cellcan be divided into three zones (Fig. 6)(13):Zone l: near the graphite, has the highest siliconconcentration and the lowest manganeseconcentration.Zone lll: at the solidification cell boundary, has thehighest manganese concentration and the lowestsilicon concentration.Zone ll: represents the rest of the matrix and hasintermediate silicon and manganese concentrations.Since the first stage incubation and reaction times arefunctions of the composition of the initial austenitephase at high temperature, it is clear that the end of thefirst stage reaction is not the same in Zones l, ll and lll.Fig. 7 illustrates, for example, the case of theapparently finished first stage reaction followed by awater quench operation. ln this case, it is clear thatZone lll will be embrittled by martensite. Fig. grepresents the different microstructures observed after5, 10, 15 and 30 min and 20 h at 380.C.Under these conditions it is very difficult to determinewhat types of austenite compose the austenite phasepercentage (Fig. 4c). ln particular, this austenite phasecontains various austenites presenting different Mgtemperatures above and below 20.C.

Fig. 6 - Repartition of various alloying elements inside the solidification cell {Zones l,ll,llt)

,q,.

Fig. 7 - Belation between the varlation of composition in Zone ll and Zone lll and theformation 0f rnartensite ln Zone lll after an apparently finished first stage

--l-rII

------\--\\

b

\_

: - ---l\l

<-+I

water quench log t

-

Zone ll------- Zonellla: beginning ofthe 1s'stage reactionb: end of 1'lstage reaction

It wìll be seen from Fig. 8 that the distribution of thevarious phases is very heterogeneous. For example, inFig. Ba, the f irst stage bainitic reactìon has started intoZones I and ll, but some areas situated in Zone lll aref ree af any bainitic products.A comparison of microstructures obtained after differentbaìnitic reaction times shows very clearly the evoluttonof this microstructural heterogeneity. Although the

Fig. B - Evolution of the microslructure as a function of the austempering time al

3BO'C:-a- t: Smin-b- r:15min-c- t:30min-d- t:20h

M etal I u rgi cal Scr e nce a nd Technology

fundamental mechanisms involved during the secondstage reactions differ from those occurring during thefirst stage reaction, the microstructural heterogeneitypersists (14).For these reasons, in ductile irons (S.G. cast irons), it isvery difficult to obtain a matrix representing only a first-stage-reaction microstructure.Under these conditions, it is clear that hardnessmeasurements (Fig. 4a) do not represent a uniquestructure, and only the evolution can be analysed.lf the best abrasive wear resistance is obtained in aspecimen which exhibits the lowest hardness value,and the highest retained austenite percentage, thismeans that the phases composing the microstructuremust be:- martensite cr'

- supersaturated ferrite (a)

- initial austenite y- enriched austenite (y)

- austenite " y ", partially enriched.It is not possible to associate a better wear resistancewith the presence of martensite which is a very hardphase. This applies also to ferrite and silicon carbidesobtained affer 20 h at 380'C.Fig. 9 shows the average gradient of each curvedescribed in Fig. 3, as a function of the bainitic holdingtime at 380"C compared with the retained austenitepercentage. There is no doubt about the choice of thepreliminary bainitic heat treatment. The problemremains, however, as described above. The austeniticphase responsible for this improved wear resistance isa mixture of austenites. So, it was interesting to studythe stability of these austenites during an abiasion test.This study was done by observing the magneticproperties, measured after every 500,000 cycles. Theresults are plotted in Fig. 10. During the abrasion test,the austenite percentage varies and gradually

Fig. I - Relation between the retained austenite percentage and the averagegradìent of different wear rate curves as a function of the auslempering time

Retained austenite %25

22.5

20

17.5

12.5

10

7.5

5

2.5

0 1 s i0 15 30 12oo min Holding time at 380 .C

a _ ^y%o

a

-

A(Ap/p) 106.t (hours),1

0.14

0

Vol. 5 [s] (1987)

10-6Tesla .m3.kg-1

120

110

100

90

80

70

60

Abrasion time- o initial- 40 hours- 80 hours- 120 hours

1 5 10 15 30 l2oominHoldingtimeat3S0.C

Fì9. 10 - Variation of "o" as a function of abrasion time

decreases, but the most spectacular result is that thisevolution is strongly related to the initial bainiticreaction holding time.The retained austenite change during wear testing isextremely important. ln fact, it can be seen that thischange becomes more important if the bainitic holdingtime is shorter (about 3%). This can be explained by theassociation of two mechanisms:- relaxation of internal stresses in the austeniticphase;

- application of external stresses to the relaxedaustenitic phase.

After a martensitic transformation occurring in ahomogeneous matrix quenched in a medium at atemperature situated in the Ms-Me scale, the structurecontains martensite and austenite. lt is known thatrelaxation of stresses within the austenite can promotea new formation of martensite. This can be done eitherby tempering at 100'C and cooling at room temperature(15) or thinning the austenitic matrix with an Mgtemperature situated below room temperature andcontaining stresses (1 1).Similarly, it is also well known that an applied stresspromotes the progression of martensite reactionswithin an apparently stable martensito - austeniticstructure. ln this study it is evident that the retainedaustenite, containing several types of austenite will nothave the same response to this new martensiticformation.The above conditions are present only at the start of thebainitic first stage reaction in Zones I, ll and lll. Themajor portion is encountered after rather short bainiticholding times, and the phenomenon is very noticeable(t < 15 min). With a longer bainitic reaction time thevariation becomes less and less important.Stress relaxation is obtained by raising the temperatureat the surface of the test specimen (60-80'C) and theapplication of stress is achieved by abrasive particlesimpinging on the abraded surface. At this temperature

it is evident that only a martensitìc transformation canbe observed.Under these conditions, the production of newmartensitic particles certainly accelerates the processof loss of matter, but after a certain tìme, equtltbrium isachieved withìn the surface Iayer. In all cases theenrìched austenite remains untransformed.Fig. 1 1 shows the appearance of the surface afterabraslve wear on a specimen austempered for 5 min. ltwìll be clearly seen that the surface does not responduniformly to the abrasive process. Near the hole(representing the site of a graphite nodule) the matrix iscomposed of supersaturated ferrite (a) and enrichedaustenite (y). The rest of the matrix consists, for themost part, of martensite s' and austenite y containingparticles of (a)and some areas of (y).

The damage caused by abrasive particles is moreserious in Zones ll and lll (grey matrrx) than in Zone I

(whìte matrix) near the graphite area. ln all cases, it isthe " centred " phase (tetragonal or cubic)which tsextracted f rom the austenitic matrix and the role of thisphase seems to be most important (16-17). lnparticular, Fig.12 indicates thìs loss of martensitic orferritic particles making it possrble to recognize theclassical " feathery " upper barnitic structure.

Fig. 11 - (SEM) analysis of an austempered specimen austempering time t : 5 min

.' 'F sop

50;rm30J kW 4.06E2

Metallurgical Sctence and Technology

;t".

&eI

Fig. 1 2 - (SEN/) analysis of the malrix after extraction of fenìte and martensiteparticles.

Conclusions

This study has demonstrated the importance of theupper barnitic structure during abrasion test at roomtemperature.The lowest wear rate is associated with the highestvolume percentage of retained austenite.This critical volume percentage of retained austeniteis not related to the apparent or the real end of thefirst stage baìnitìc reaction.The low hardness associated with this high retainedaustenite percentage corresponds to a high wearabrasion resistance.Formation of martensite is observed during the testat and near the abraded surfaces.The wear is not uniform in the matrix. lt is a f unctionof the solidification cell structure which isheterogeneous in compositìon.

REFERENCES

(1) Gundlach, R.8., and J.F. Janowak. A review of austemperedductile iron metallurgy. ln ASM (Ed.), First lnternational ionferenceon Austempered Ductile lron: Your Means to ImprovedP^erformance, Productivity and Cost, Hyatt Regency O,Hare,Chicago. lll., 24 Aprit 1984. ASM, Metats park. Òhio, tgAq, pp. 1-12.

Q) Watmough, T., and M.J. Malatesta. Strengthening of ductile ironfor crankshaft applications. A.F.S. Trans., 92119A), B3-gg.(3) Johansson, M., A. Vesanen, and H. Rettig. Austenitic-bainiticnodular cast tron as a construction material fòr gears.Antriebstechnik, 15(1 1 ) (1976), 593-600 (BCIRA transtation Tt 507).(4) Dodd, J. High-strength, high-ductitity, ductile irons. ModernCasting, 68 (1 978), 60-66.(5) Zaccone, M.A., and D.E. Bowman. Applications of austemperedductile iron in the heavy equipment industry. ln ASM (Ed.), Firàtlnternational Conference on Austempered Ductite lron: yourMeans to lmproved Performance, Productivity and Cost, HyattRegency O'Hare, Chicago, lll., 2-4 April 1984. ASM, Metats park,Ohio, 1 984, pp. 227-252.(6) Corso, S. Development of bainittc nodular iron for theconstruction of speed gears for the car industry. EEC CommissionReport EUR 8639 (Final report). Commission of the EuropeanCommunities, Luxembourg, 1983, (in ltalian, with extended Englishsummary).

(7) Forrest, R.D. Austempered ductìle iron for both strength andtoughness. Machine Design, 57 (22) (t 9BS), 95-99.

(8) Gagne, M., and P.A. Fallon. Microstructural characteristics ofbainitic ductile irons. Twenty-third Annual Conference ofMetallurgists, Quebec City, August 1984.

(9) Harding, B.A. Prospects for the exploitation of austemperedductile irons. Presented at the Second lnternational Confetrence onAustempered Ductile lron, March 1986, Ann Arbor, Mi.(10) Schissler, J.M., and J. Saverna. Effect of the pretiminarystructure of steel 230C13 upon the improvement of resistance to hotabrasion. ln ASME (Ed.), Proceed. of Wear 1983, Reston-Washi ngton, pp. 474-483.

(1 1) Schissler, J.M. Study of the bainitic transformation, at 420"C, inFe-C-Si alloys. Thesis, Nancy, 1972.

(12) Schissler, J.M., B. Vigneron, and R. Faivre. The bainitictransformation in Fe-C-Si alloys. Hommes et Fonderie, 78 (1977), 17-34.

(13) Schissler, J.M., and C. Bak. Scanning electron microscopeexamination of the embrittlement of ferritic ductile cast iron. A.F.S.Trans., 88 (1 980), 301-31 1 .

(14) Schissler, J.M., J. Saverna, R. Bettocci, and C. Bak.S.G. cast iron influence of hot shakeout on the microstructureoltained through varying heat treatments. ln physicat Metallurgy ofCast lron. North-Holland, Amsterdam, pp. 335-344.(l5) Simon, A. Study of the tempering of martensite in Fe-C atloys.Thesis, Nancy, 1972.

(16) Voigt, R.C., H. Dhane, and L. Etdoky. Microstructural aspects offracture in austempered ductile iron. ln ASM (Ed.), Proceed. of theSecond lnternational Conference on Austempered Ductile Iron,Ann Arbor, Mi., 1986.

(17) Vetters, H., P Mayr, and J. Walla. lnvestigations on the stressinduced martensite formation in ductile iron. lbid.

Vot. 5 t3l(1987)